Evaluating the performance of machine learning models for energy load prediction in residential HVAC systems

Given that heating and cooling systems account for roughly 40% of overall building energy usage, accurately predicting and analyzing the energy needed for HVAC operations in residential buildings is crucial. This research proposes the application of machine learning algorithms to precisely forecast the cooling and heating loads of residential buildings, enabling a practical advancement towards the design of energy-efficient and sustainable building structures. The study leverages a comprehensive dataset derived from the National Renewable Energy Laboratory (NREL) ResStock data, encompassing 550,000 data points representing diverse residential building characteristics, occupant behaviors, and weather conditions across the contiguous United States, providing a broad spectrum of geographic regions and climate zones. To evaluate the effectiveness of various machine learning models for residential energy prediction, we compared the performance of linear regression, random forest, support vector regression, gradient boosting, decision tree, neural network, and Convolutional Neural Network - Long Short-Term Memory (CNN-LSTM) across key metrics such as R-squared, Root Mean Squared Error (RMSE), and Mean Absolute Error (MAE) for both cooling and heating loads. Our findings indicate that neural networks consistently achieved the highest accuracy, demonstrating superior prediction capabilities during both training and testing phases. Notably, gradient boosting emerged as a strong contender due to its balance between prediction accuracy and computational efficiency, making it particularly suitable for resource-constrained environments. These results highlight the potential of machine learning to provide actionable insights for stakeholders in residential buildings, enabling informed energy management strategies aimed at reducing consumption and mitigating climate change.